|Publication number||US5123085 A|
|Application number||US 07/495,338|
|Publication date||16 Jun 1992|
|Filing date||19 Mar 1990|
|Priority date||19 Mar 1990|
|Also published as||DE69131927D1, DE69131927T2, EP0521084A1, EP0521084A4, EP0521084B1, WO1991014995A1|
|Publication number||07495338, 495338, US 5123085 A, US 5123085A, US-A-5123085, US5123085 A, US5123085A|
|Inventors||Stuart C. Wells, James V. Loo, Dawn M. Wallner|
|Original Assignee||Sun Microsystems, Inc.|
|Export Citation||BiBTeX, EndNote, RefMan|
|Patent Citations (8), Referenced by (159), Classifications (10), Legal Events (8)|
|External Links: USPTO, USPTO Assignment, Espacenet|
1. Field of the Invention
This invention relates to the rendering of graphic images on computer displays. More specifically, this invention relates to the rendering of anti-aliased polygons on computer graphic displays.
2. Art Background
In the scan conversion process of a polygon, the polygon, which is defined by the coordinates of its vertices, is converted to a raster image. The outline or edges of the polygon are first determined and the pixels between the polygon edges are then filled. The vertices of the polygon are translated into the display coordinate space and a line scan conversion algorithm is used to determine the lines between vertices which form the edges of the polygon. The line scan conversion algorithm computes the coordinates of the pixels which lie closest to the line between vertices on a two-dimensional raster grid. An example is Bresenham's line algorithm (for information on Bresenham's line algorithm see: Foley and Van Dam, Fundamentals of Interactive Computer Graphics, (Addison-Wesley Publishing, 1984) pages 433-436). However, the lines do not always coincide or "line up" with pixel locations. Although the closest pixel is chosen, the resulting line contains the undesirable alias effect. This undesirable effect also is carried into the polygon rendering process.
In an attempt to smooth out the edges of a polygon during the polygon rendering process, Pitteway and Watkinson modified Bresenham's algorithm to determine a partial coverage value for pixels which form the edge of a polygon. The modification involved adding to Bresenham's error term the quantity (1-m), where m is the slope. The result is that the error term is a measure of the area of the pixel inside the polygon (for further information on the Pitteway/Watkinson algorithm see Foley and Van Dam, Fundamentals of Interactive Computer Graphics, (Addison-Wesley Publishing, 1984), pages 460-463). Using the Pitteway/Watkinson algorithm, a pixel intensity is generated which is equal to the pixel area covered multiplied by the color or intensity of the polygon at that pixel (since a shaded polygon may result in a different intensity or color).
The coverage value computed using the Pitteway/Watkinson algorithm is the area between the line of the polygon and the major axis. If the polygon lies toward the major axis, i.e., between the line being drawn and the major axis, the coverage value can be used directly. If the polygon lies toward the minor axis, the desired coverage value equals the difference between the value of one and the Pitteway/Watkins coverage value determined. Because the Pitteway and Watkinson algorithm can only address a single line, the algorithm cannot handle thin polygons whose thickness may be less than one pixel and which may reside within the area of a single pixel, or vertices which meet within the area of a single pixel. In addition, since the algorithm touches one pixel per major axis increment, some pixels are partially covered by a polygon are treated as though they are uncovered. Furthermore, the Pitteway/Watkinson algorithm only performs a simple averaging process and does not extend direclty to different convolution or weighted averaging functions.
Another approach to rendering a polygon employs a sub-pixel mask which indicates the locations of coverage within the pixels. For example, a sub-pixel mask may provide a 16-bit pattern on a 4×4 grid of coverage for a polygon on a pixel. The mask represents coverage of combination of 16 sub-pixel areas plus full coverage and no coverage which provides for 256 different combinations of sub-pixel masks. To accommodate horizontal and vertical edges lying on integer pixel boundaries, the look-up table indices access a 5×5 array. This is illustrated by FIGS. 1a, 1b and 1c. The mask obtained from the table is a 4×4 array and is composed of bits where a value of zero indicates that the polygon does not cover the sub-pixel and a value of 1 indicates that it does. To use the sub-pixel mask, the polygon vertices are quantized to sub-pixel values after passing through any geometric manipulations. To calculate the indices for mask look-up the sign of the edge of the slope must be provided. This determines the integer x coordinate of the pixel which is adjacent to the current pixel in the direction of the slope. By convention, horizontal edges are rendered left to right so the adjacent pixel is always to the right. Vertical edges require the additional knowledge of whether they are left or right polygon edges to determine adjacent pixels. Referring to FIG. 1a, when the adjacent pixel's x value is subtracted from the current x value of the edge of the current pixel and the result is multiplied by the number of sub-pixels in x (4 for the present example) the following results are determined:
0=0 * 4;
1=0.25 * 4;
2=0.50 * 4;
3=0.75 * 4;
4=1 * 4
These values become the x index into a sub-pixel identifying matrix.
FIG. 1b illustrates the xy index and FIG. 1c depicts the sub-pixel identifying index. The same calculation performed for x is performed for the y values of the edge. The calculations to determine the y indices are simplified by the fact that the polygon is rendered a scan line at a time in the increasing y direction. The x and y indices give us sub-pixel number. (e.g. 0 through 24) of the entry index of the pixel. Sub-pixel masks are designed to allow the entry index to be within the pixel as well as on a pixel boundary. Identical procedures are performed for the exit sub-pixel. The exit sub-pixel is calculated as soon as an adjacent pixel or scan line has been intercepted or when the edges ends (if it ends within the current pixel).
Once the two sub-pixel indices for the entry and exit sub-pixels are determined, the mask is obtained from the look-up table and is either used as is or complemented depending upon the edge code which indicates which side of the edge is an inside edge and which is an outside edge. The mask for multiple edges of a single polygon within a pixel are combined depending upon a vertex code; the masks are logically ORed together if the vertex represents two edges whose inside angle is greater than 180° (concave) and logically ANDed together if the inside angle is less than or equal to 180° (convex) The sub pixel masks themselves may be determined empirically or according to a predetermined algorithm so long as consistent algorithms are used to generate all the masks. Memory may be saved by providing only half of the possible combinations of edge crossings in as much as the other half may be obtained by complementing the bit pattern based upon the knowledge of the relationship of the edges in a mask to the inside/outside of the polygon.
Once the desired partial coverage pixels on the edge of a polygon have been computed, the internal pixels between the polygon edge pixels on a span on a scan line must be filled. Filling a span on a scan line requires determination of the right most edge pixel on the left edge of a polygon and left most edge pixel on the next right edge of the polygon. Once determined, the pixels between these edge pixels are filled. FIG. 2a illustrates why it is necessary to determine the first and last fill pixels in a span since simply using the extreme edge pixels of the polygon on a span would result in adjacent edge pixels being entirely painted as opposed to partially painted i.e. "color* coverage" for edge pixels which are partially covered and simply "color" for fill pixels which are fully painted. FIG. 2b shows an example where only two pixels in the span should be entirely filled since there are five left edge pixels (including one vertex) and two right edge pixels.
For a single polygon there may be several spans on an individual scan line which must be filled (See Foley and Van Dam, Fundamentals of Interactive Computer Graphics, p. 456, for a description of rendering a polygon). Determining the individual spans on a particular scan line involves rendering all edges of the polygon on that scan line and determining the first and last pixels in a span to be completely filled. Upon determination of these first and last fill pixels, an X order sort is performed on their position in the scan line and individual spans are then filled in a left to right order.
Once the energy or coverage for each pixel of a polygon is determined, the polygon image may then be composited into the background. Compositing is a technique wherein the pixels of the polygon are blended into the background in order to have as smooth a transition as possible between the polygon and the background. In one method for compositing, a separate component or channel, referred to as the alpha "α" channel, is used to identify the extent of coverage of an element at a pixel. See Porter & Duff, "Compositing Digital Images", Computer Graphics, Vol. 18, No. 3, (July 1984), In order to place the element over an arbitrary background, a mixing factor is required at each pixel to control the linear interpolation of the foreground and background colors. This information is encoded into the α channel, which at a value of zero indicates no coverage and at a value of 1 means full coverage with fractions corresponding to partial coverage. The elements of an image (pixel) are then composited into the background according to the following compositing equation:
c.sub.t =α.sub.a *c.sub.a +(1-α.sub.a)*α.sub.b *c.sub.b
α.sub.t =α.sub.a +(1-α.sub.a)*α.sub.b
where ca is the color of the pixel of the polygon, cb is the color of the background object, ct is the color of the pixel upon compositing, αa is the value of the partial coverage of the pixel of the polygon, αb is a partial coverage of the background pixel and αt is the composited α value.
In one implementation of the α channel the α channel retains only a single α value per pixel which is proportional to the partial coverage at a pixel by the polygon. No information associated with the orientation or shape of the partial coverage is retained. In addition, if the number of polygon edges intersecting a pixel is greater than 1, the α value becomes a composite α value, and the actual number of intersecting polygons is unknown. These compositing equations are equally applied to all pixels of the polygon. Therefore, if the polygon ranges in depth, the pixels of the polygon will be rendered the same regardles of depth. Furthermore, the pixel color computed via the compositing equation is heavily dependent upon the rendering sequence of the polygons which overlap at a particular pixel. The closer the rendering sequence is to the actual depth order of the polygons, the more accurate is the computed color value to the desired color value. U.S Pat. No. 4,748,572 describes a mechanism in which the polygons to be rendered are sorted in depth order prior to compositing. However, the sorting process is quite time consuming, requires several processing passes and slows down the polygon rendering process considerably.
The error which arises when rendering intersecting polygons and which varies according to the depth order of the polygons is illustrated by the example of FIGS. 3a and 3b. FIG. 3a shows a pixel within which three polygons a, b, and c intersect. Each pixel within a polygon has a depth value associated with it identified as the z value and an α value indicative of the extent of coverage at the pixel. FIG. 3b shows the resulting color and α values which have been computed for each polygon for all possible depth orders for αa equal to one-half, αb equal to one-half and αc equal to one-half. FIG. 3b shows that in 4 out of 6 cases considered, cases 1, 2, 4 and 6, the contributions to the final pixel color from the three intersecting polygons are dependent upon the depth order. The error that arises as a result of this is referred to as the depth order dependence error.
FIG. 4 illustrates two cases of polygons a, b and c, having, αa equal to one-half, αb equal to one-quarter and αc equal to one. The sequence of rendering is c, b, a. It should be noted that when α obtains the value of one, it always remains at the value of one regardless of subsequent compositing with other polygons. FIG. 4 verifies this to be true irrespective of depth order. In addition, in the second case, there is no contribution to the final color from polygon "a" despite polygon a being in front of polygon c and having an α value equal to one-half. However, because the z value is updated after the computation of the color to a value of z=zb (at the second step), and the α values is equal to one, there is insufficient information to conclude anything other that polygon a is hidden by a polygon completely covering the pixel (even though the other polygon is actually in front of the polygon having an α value of 1).
It is therefore an object of the present invention to generate high quality polygons which are rendered in a random order independent of the polygon depth.
It is an object of the present invention to render high quality polygons utilizing a single series or single pass of calculations which renders the polygon in the minimum amount of time.
It is an object of the present invention to render anti-aliased polygons wherein, on a per pixel basis, the z value of each pixel is determinative of the compositing equations used to composite each polygon pixel with the corresponding background pixel.
It is further an object of the present invention to provide a polygon rendering process wherein a sub-pixel mask for each pixel is used in conjunction with a depth value to generate anti-aliased polygons.
A scan conversion process is performed on a polygon using a single pass technique. The pixels which comprise the edges and vertices of the polygon are first determined from the vertices which defines the polygon. The alpha channel comprises either a sub-pixel mask associated with each pixel which indicates the amount and sub-pixel regions of coverage or a single value indicative of the percentage of coverage of a pixel. Furthermore, a z value indicative of the depth of each pixel is maintained. The pixels between the edge pixels of the polygon are then turned on, thereby filling the polygon. The pixels which comprise the polygon are then composited with the background pixels on a per pixel basis. The depth value of each pixel of the polygon (the z value) is used to determine the compositing equations to be used to composite each pixel of the polygon to the background. The compositing equations update the color of the pixel, the z buffer value of the background pixel and the sub-pixel mask to reflect the addition of information from the compositing of the pixel of the polygon into the background pixel. Through this method high quality anti-aliased polygons may be rendered without performing the time consuming process of sorting the polygons in depth order prior to compositing.
In one embodiment, a sub-pixel mask registration, which describes the coverage for each pixel of the polygon and the background, is maintained. A polygon is rendered by first determining the edges of the polygon using a scan line conversion algorithm and determining the sub-pixel mask registration for the edge pixels of the polygon. The polygon is then filled by filling in the pixels between edge pixels. Once the polygon is filled, the polygon is composited pixel by pixel with the background using a set of equations which are dependent upon the polygon's z value with respect to the background pixel's z buffer value. Sub-pixel masks are maintained for the background pixels and are updated according to the compositing equations. The z buffer value is also determinative of whether the color is modified for a particular pixel, whether the z buffer value is modified for a particular pixel and whether the sub-pixel mask (determinative of the coverage of the pixel) is modified. Thus, a highly accurate rendering of an anti-aliased polygon is achieved without the cost or speed required to sort the polygons to be rendered into z buffer order.
In another embodiment, to minimize the amount of memory required, a single value, the alpha value, is used to indicate the amount of coverage of the background pixels. Prior to compositing, the sub-pixel masks of the polygon are converted to a single value. A single value indicative of coverage is also maintained for each of the background pixels. The z value of each pixel of the polygon to be composited is compared with the corresponding z buffer value of the background pixel. The polygon pixel is then composited into the background pixels using compositing equations dependent upon the comparison of the z value and z buffer value.
The objects, features and advantages of the present invention will be apparent from the following detailed description in which:
FIGS. 1a, 1b and 1c illustrate the concept of a sub-pixel mask for each pixel.
FIGS. 2a and 2b illustrate the definition of the first and last pixels filled between edges of a polygon.
FIGS. 3a and 3b illustrate the result of compositing random depth order polygons.
FIG. 4 is a table showing the results of compositing in a random depth order.
FIG. 5 is a block diagram illustration of a computer which employs the present invention.
FIG. 6 is a flow chart which generally depicts the preferred embodiment of the present invention.
FIG. 7 is a diagram which illustrates the Pitteway/Watkinson scan conversion algorithm.
FIG. 8 illustrates the determination of coverage of thin polygons within a pixel.
FIG. 9 illustrates an eight bit sub-mask.
FIGS. 10a, 10b and 10c illustrate an exemplary polygon edge and corresponding inclusive and exclusive sub-pixel masks.
FIGS. 11a, 11b, 11c, 11d11e, 11f, 11g, 11h, 11i, and 11; illustrate the formulation of alpha values and sub-pixel masks.
FIG. 12 illustrates the use of a look-up table to determine the sub-pixel mask for a pixel.
FIGS. 13a-13n illustrate an example of the generation of sub-pixel masks for a polygon.
FIGS. 14 illustrates a polygon used to describe a scan line conversion algorithm.
FIGS. 15a and 15b illustrate the compositing equations employed.
FIGS. 16a, 16b, 16c, 16d, 16e, 16f, 16g, 16h and 16i illustrate the depth order dependence error and improvements to the same through the system of the present invention.
FIGS. 17a, 17b and 17c further illustrate the depth order dependence error and improvements to the same through the system of the present invention.
The detailed descriptions which follow are presented largely in terms of algorithms and symbolic representations of operations on data bits within a computer memory. These algorithmic descriptions and representations are the means used by those skilled in the data processing arts to most effectively convey the substance of their work to others skilled in the art.
An algorithm is here, and generally, conceived to be a self-consistent sequence of steps leading to a desired result. These steps are those requiring physical manipulations of physical quantities. Usually, though not necessarily, these quantities take the form of electrical or magnetic signals capable of being stored, transferred, combined, compared, and otherwise manipulated. It proves convenient at times, principally for reasons of common usage, to refer to these signals as bits, values, elements, symbols, characters, terms, numbers, or the like. It should be borne in mind, however, that all of these and similar terms are to be associated with the appropriate physical quantities and are merely convenient labels applied to these quantities.
Further, the manipulations performed are often referred to in terms, such as adding or comparing, which are commonly associated with mental operations performed by a human operator. No such capability of a human operator is necessary, or desirable in most cases, in any of the operations described herein which form part of the present invention; the operations are machine operations. Useful machines for performing the operations of the present invention include general purpose digital computers or other similar devices. In all cases there should be borne in mind the distinction between the method operations in operating a computer and the method of computation itself. The present invention relates to method steps for operating a computer in processing electrical or other (e.g. mechanical, chemical) physical signals to generate other desired physical signals.
The present invention also relates to apparatus for performing these operations. This apparatus may be specially constructed for the required purposes or it may comprise a general purpose computer as selectively activated or reconfigured by a computer program stored in the computer. The algorithms presented herein are not inherently related to a particular computer or other apparatus. In particular, various general purpose machines may be used with programs written in accordance with the teachings herein, or it may prove more convenient to construct more specialized apparatus to perform the required method steps. The required structure for a variety of these machines will appear from the description given below.
FIG. 5 shows a typical computer-based system for rendering anti-aliased polygons according to the present invention. Shown there is a computer 101 which comprises three major components. The first of these is the input/output (I/O) circuit 102 which is used to communicate information in appropriately structured form to and from the other parts of the computer 101. Also shown as a part of computer 101 is the central processing unit (CPU) 103 and memory 104. These latter two elements are those typically found in most general purpose computers and almost all special purpose computers. In fact, the several elements contained within computer 101 are intended to be representative of this broad category of data processors. Particular examples of suitable data processors to fill the role of computer 101 include machines manufactured by Sun Microsystems, Inc., Mountain View, Calif. Other computers having like capabilities may of course be adapted in a straightforward manner to perform the functions described below.
Also shown in FIG. 5 is an input device 105, shown in typical embodiment as a keyboard. It should be understood, however, that the input device may actually be a card reader, magnetic or paper tape reader, or other well-known input device (including, of course, another computer). A mass memory device 106 is coupled to the I/O circuit 102 and provides additional storage capability for the computer 101. The mass memory may include other programs and the like and may take the form of a magnetic or paper tape reader or other well known device. It will be appreciated that the data retained within mass memory 106, may, in appropriate cases, be incorporated in standard fashion into computer 101 as part of memory 104.
In addition, a display monitor 107 is illustrated which is used to display messages or other communications to the user. Such a display monitor may take the form of any of several well-known varieties of CRT displays. Preferably, the display monitor 107 may also display the graphic images, i.e., the anti-aliased polygons rendered according to the process of the present invention. A cursor control 108 is used to select command modes and edit the input data, and in general provides a more convenient means to input information into the system.
Through the present invention, high quality anti-aliased polygons may be rendered in a single pass without performing the time consuming process of sorting the polygons in depth order. Each polygon to be rendered is identified by its vertices, its color or intensity and its depth value, also referred to as the z buffer value.
Referring to FIG. 6, at block 300 the coverage of the pixels which form the edges of the polygon to be rendered are determined. The amount of coverage is reflected in the intensity of the pixel, the greater the coverage, the greater the contribution of the polygon pixel to the pixel's final intensity. The coverage value is referred to as the α (alpha) value. Preferably, the α value is determined according to the Pitteway/Watkinson algorithm. The coverage value computed according to the Pitteway/Watkinson algorithm is the area between the line representing the edge of the polygon and the major axis. This is further explained by reference to FIG. 7. The desired line to be rendered is identified by numeral 400. Bresenham's line algorithm determines which pixel, for example the pixel ti 410 or the pixel si 420, to render to best estimate that portion of the line. The algorithm uses a decision variable di which is proportional to the difference between s and t. Pi-1 has been determined to be the closest pixel to the actual line being drawn and now di is computed to determine whether the next pixel to be set should be ti or si. If s is less than t, then si is closer to the desired line and should be set, else ti is closer and should be set. This may be looked at from different perspective by noting that si is chosen if s-t is less than zero. The d value is computed to be 2* (dy-dx) where dx equals the absolute value of (x2-x1) and dry equals the absolute value of (y2-y1 ).
Using the Pitteway/Watkinson algorithm, the d value is computed to be equal to (mx-y)+(1-m), where m is the slope of the line. The test to determine which pixel is to represent that portion of the line does not compare the value to zero as in the original algorithm but rather to the value (1-m). The variable d then serves to estimate the pixel coverage at each step. The closer d is to 1, the more of the pixel is covered and the greater the contribution to the pixel intensity. However, the Pitteway/Watkinson algorithm does not adapt well for thin polygons or vertices.
In order to minimize the error which occurs when attempting to render thin polygons a technique is employed where the edges of the polygon are individually rendered into an intermediate scan line buffer. This technique is preferred because an edge is generally rendered fully for a particular scan line and partial coverage values for a single edge must be stored for subsequent computations. For example, it is necessary to store the partial coverage mask for a single edge because the coverage is the combination of the masks obtained from traversing two edges (for example, the two edges illustrated in FIG. 8). The edge pixels of the polygon are first sorted according to y coordinate values such that edge pixels located on the same scan line may be rendered sequentially and therefore together into the intermediate scan line buffer. Thus, the coverage for a thin polygon or vertex may be individually determined prior to compositing into the background.
Referring to FIG. 8 the final coverage of the thin polygon or vertex may be derived using the following compositing equation: ##EQU1## where αa is the partial coverage between the lower polygon edge and the upper scan line and αb is the partial coverage between the upper polygon edge and the scan line. Similarly, if sub-pixel masks are used, the scan line buffer would consist of a scan line to the resolution of sub-pixel values. Thus, the final coverage of the thin polygon or vertex may be derived using the following compositing equation:
M.sub.F =M.sub.A AND M.sub.B
where "AND" represents the logical AND operation, MF represents a full sub-pixel mask, MA represents the sub-pixel mask for the lower polygon edge and MB represents the mask for the upper polygon edge.
The coverage of each of the edge pixels of a polygon may be represented by the α value or sub-pixel masks. The sub-pixel masks divide the area of a pixel into sub-pixel areas. The sub-pixel areas may be of any size or shape, although it is preferred that the sub-pixel areas be equal in area. In the present embodiment, the sub-pixel masks are designed to be equal in area by dividing the pixel area according to a 4×4 grid to provide a 16bit pattern of coverage for a polygon on a pixel. The mask represents coverage of a combination of 16 sub-pixel areas plus full coverage and no coverage. Each sub-pixel element represents a geographic location of a sub-pixel. Thus, the mask provide a more accurate representation of the coverage of the pixel because not only the amount of coverage but the actual location of coverage within a pixel is determined. Alternatively, to minimize the number of masks in the look-up table stored in memory, an eight bit mask may be used in which the pixel area is divided into eight sub-pixel areas according to a predetermined or randomly generated pattern such as the mask illustrated in FIG. 9.
The set of masks in the look-up table stored in memory should provide a sub-pixel mask for all possible combinations that may occur. To save space in memory, it is preferred that the number of masks stored be reduced by one-half by not storing those masks which are the complement of or are a mirror image of a mask already stored. During the rendering process, those masks not stored may be easily determined by complementing the mirror image mask stored in memory.
The masks may be generated empirically or through a predetermined process which addresses all possible combinations of masks. The procedure to generate the masks should insure that actual amount of coverage area is approximately equal to the amount of coverage area determined for the corresponding mask. This is illustrated by FIGS. 10a and 10b. FIG. 10a shows an edge through a pixel wherein the inclusive side of the edge covers approximately 5/8 of the area of the pixel. The corresponding mask, shown in FIG. 10b has sub-pixels set that cover 9/16 of the area of the pixel. This mask accurately represents the edge because the areas of coverage are approximately equal.
The mask generation procedure should also be able to generate the mask for the area on one side of the edge and the mask for the other side of the edge whereby if the two masks are added together, the resultant mask equals full coverage. This is illustrated by FIGS. 10b and 10c. FIG. 10b, as stated earlier, represents the corresponding mask generated for the inclusive side of the edge illustrated in FIG. 10a. FIG. 10c is the mask generated for the exclusive side of the edge illustrated in FIG. 10a. The sum of the masks in FIGS. 10b and 10c result in a full mask which indicates that the process for the generation of the masks is accurate.
One way to generate the masks is by determining a set of masks for each sub-pixel which correspond to the areas drawn by a plurality of vectors through the center of the sub-pixel which span the range of angles. The vector may be rotated by a predetermined increment and the mask computed according to a scan conversion process which determines the edge sub-pixels (such as Bresenham's) and filling in all sub-pixels within the area delineated by the edge sub-pixels. This process is performed with respect to each sub-pixel.
Another method to generate the masks is by using the Pitteway/Watkinson algorithm to determine the masks according to a predetermined area size and entry point. The area size and entry point is then varied to generate the different masks.
An illustration of the formulation of alpha values and corresponding sub-pixel masks are provided by FIGS. 11a, 11b, 11c, 11d, 11e, 11f, 11g, 11h, 11i, and 11j. In this example, 16 cell tables are addressed by two intersection points. It is assumed that the intersection point for the left edge is (0.0, 0.0) and that the facet to be represented resides to the right of this left edge. To calculate the sub-pixel masks and coverage values, the Bresenham line algorithm is employed to transverse the display element. However, the step size is 0.25 rather than 1.0.
FIGS. 11a, 11b, 11c, 11d, 11e, 11f, 11g, 11h, 11i present the masks for the different combinations of intersection points for facets to the left of the edge. The facets to the right of the edge can be determined simply by complementing the corresponding mask for the facet to the left of the edge.
FIG. 11j presents a table of alpha or coverage values for facets to the left of the edge. These may be determined simply by summing the number of sub-pixels set in the mask. A value of 16 represents full coverage. The coverage values for facets to the right of the edge may be determined by subtracting one coverage value for the facet to the left of the edge from the value of sixteen.
The masks are incorporated into an array or set of tables which are stored in memory. During the rendering of anti-aliased polygons, the masks are accessed according to an index determined from the edge which resides within the pixel. The index may be the entry sub-pixel and exit sub-pixel, that is the sub-pixels at which the edge enters the pixel and exits the pixel. Alternatively, the index may consist of the entry sub-pixel and the slope of the edge through the pixel. The index may also be the slope of the edge through the pixel and the perpendicular distance from the center of the pixel to the edge. In addition, for the case of determining the mask for the vertex which occurs within a pixel, the index may be the entry sub-pixel and the sub-pixel location of the vertex.
Referring again to FIG. 1, it is preferred that a 5×5 table is used to represent a 4×4 sub-pixel mask. The table is accessed via a sub-pixel entry index and exit index. The polygon entry and exit indices are the sub-pixel locations where the polygon edge respectively enters and exits the pixel. Polygon vertices are quantized to sub-pixel values after passing through any geometric manipulations. To calculate the entry index for the table, the sign of the slope of the edge must be determined. This is used to compute the integer x coordinate of the pixel which is adjacent to the current pixel in the direction of the slope. By convention, it is preferred that horizontal edges are rendered left to right so that the adjacent pixel is always to the right. To determine adjacent pixels of vertical edges, the additional information of whether the edge is a right or left edge is needed. The adjacent pixel x value is then substracted from the current edge pixel x value and is multiplied by the number of sub-pixels in x (4 for the present example):
The values, 0.25, 0.5, 0.75 and 1, are the quantization levels for the present mask used. These fractional values are produced because of the rounding to the nearest sub-pixel position at vertices prior to traversing the edge and subtracting the integer pixel value from the sub-pixel precision value. The value is multiplied by a value of 4 to identify the index into the sub-pixel matrix (See FIG. 1b). The same calculation is performed for the y values of the edge. These x and y indices give the sub-pixel number, in the present example 0 thorough 24, of the entry into the pixel. Masks are designed to allow the entry index to be within the pixel as well as on pixel boundaries. Identical procedures are performed for the exit sub-pixel (that is, the sub-pixel location at which the polygon edges crosses pixel boundaries). The exit sub-pixel is calculated at the location an adjacent pixel or scan line has been intercepted or where the edge ends (if it ends within the current pixel).
Once the entry and exit sub-pixel indices are determined, the mask is obtained from the look-up table. This is illustrated by FIG. 12 which shows the sub-pixel indices 509 and 510 being input to the mask look-up table 515 to derive the mask 520.
The masks generated for the situation of multiple edges occurring within a single pixel are first individually determined (that is, the mask for each edge is individually determined) and combined. In one embodiment, the masks are combined according to an edge code. The edge indicates whether the edge is a right or left edge, that is, whether the edge is convex or concave, which determines which regions are inside the edge and thus the polygon.
A pixel is inside a polygon if, at a particular scan line, the pixel is to the left of the right edge and to the right of the left edge. The identification of left and right edges may be performed a multiplicity of ways. In one method, the vertices of the polygon are sorted in descending order according to y coordinate values. The x coordinate values corresponding to adjacent y coordinate values are compared. If the x coordinate value corresponding to the greater y coordinate value is less than the x coordinate value corresponding to the lesser y coordinate value, the edge designated by the coordinates is a right edge. Conversely if the x coordinate value corresponding to the greater y coordinate value is greater than the x coordinate value corresponding to the lesser y coordinate value, the edge is designated as a left edge.
In another method, the polygon direction is determined. For example, the polygon may be described in a clockwise or a counter clockwise direction. If the polygon is described in a counter clock-wise direction and the change in x of the edges (δx=x2 -x1) is greater than zero, the edge is a left edge and the mask corresponding to the edge will be complemented. A vertical edge (δx1=0) will be coded as a left edge and be complemented if the change in y (δy1=y2-y1) is less than zero. In general edges the δy1 of which is greater than zero, are coded as left edges and those less than or equal to zero are coded as right edges. Horizontal edges proceeding from left to right (δx1>0) are coded as right edges and those proceeding from right to left are coded as left edges. Similarly, the vertex coding is accomplished using a cross product of the two participating edges' δx's and δy's calculated below:
cross product (CP)=(δx1×δy2)-(δx2×δy1)
where δx1, δy1 are from edge 1 and δx2 and δy2 are from edge 2. If the cross product is less than or equal to zero (0≧CP) the vertex is coded as convex which dictates that the two sub-pixel masks generated from the two edges are logically ANDed together. Otherwise, the vertex is concave which dictates that the two masks are logically ORed together. The vertices as well as the edges must be coded because one end of an edge at a first vertex may form a concave angle while the other end of the same edge at a second vertex may form a convex angle with another adjoining edge.
An example of the generation of pixel masks for a polygon is discussed below with reference to FIGS. 13a-13n. FIG. 13a shows a triangular shaped polygon having three vertices at (2.1, 0.2), (3.1,2.2) and (0.2, 3.1). The number to the left of the decimal point represents the pixel value and the number to the right of the decimal point represents the fractional or sub-pixel value. To generate the vertex positions, the vertex values are first translated into the display coordinate space. The vertex coordinates are then quantized to sub-pixel elements (preferably to at least two fraction bits within the pixel). The vertex positions are then sorted by Y coordinate values and by X coordinate values to establish the active edge list and the scan line order.
The scan line conversion algorithm determines that the first scan line to include a portion of the polygon is at y=0 because the first vertex is at y coordinate value 0.2. The next step is to extrapolate the sub-pixel locations of the edges (i.e., edge crossings) of the pixel because the sub-pixel locations of the edges are used to index to the proper sub-pixel mask. In the present example, the Bresenham algorithm is used to determine the intersection of the polygon edge with the edge of the pixel.
The first pixel evaluated is the pixel at (2.0, 0.0) because the first vertex is at (2.1, 0.2). This is illustrated in FIG. 13b. Because the vertex lies within a single pixel, each edge of the vertex is individually determined and combined to generate the sub-pixel mask for the vertex. Thus, as illustrated by FIG. 13c, the left edge of the vertex (the edge between (2.0, 1.0) and (2.1, 0.2)) is extrapolated to intersect the edge of the pixel using the Bresenham algorithm at sub-pixel location (2.2, 0.0). Similarly, the right edge of the vertex (the edge between (2.2, 1.0) and (2.1, 0.2)) is extrapolated to intersect the edge of the pixel at sub-pixel location (2.0, 0.0). These entry and exit points to the pixel are used to index the coverage table containing the sub-pixel masks.
The fraction bits (the value to the right of the decimal point) of the entry and exit points are extracted to create the address which forms the index into the coverage table. It should be noted that an intersection point may cross one of five sub-pixel areas: the four sub-pixels of the current pixel and the 0 sub-pixel of the next adjacent sub-pixel. Thus when an intersection point occurs on the right or bottom edge of the pixel, the more significant bit, in the present example this is the number to the left of the decimal point, is referred to. If the value of the fraction bits is zero, the more significant bit is compared to the value of the more significant bit of the current pixel. If the two values are equal, the portion of the address for that particular coordinate of the intersection point is zero. If the value of the more significant bit of the coordinate value of the intersection point is greater than the corresponding bit of the current pixel, the address for that particular coordinate of the intersection point is given a value of 4. In the present example, this process would result in the values shown in FIG. 13d. The table of FIG. 13d includes a bit, referred to as the "lr" bit, which designates whether the edge is a left or right edge. The values for each edge are utilized to address the coverage table to retrieve the corresponding sub-pixel mask for each edge.
The coverage table returns the sub-pixel mask for the infinite edge which intersects the pixel at the address determined. FIG. 13e illustrates the sub-pixel masks which would be retrieved from the coverage table. Using vertex logic, as described above, it is determined that the vertex is convex; therefore the two sub-pixel masks are combined by logically ANDing the two masks. The resultant sub-pixel mask is written into an intermediate scan line buffer which stores the pixel values for the current scan line until the scan line is completely processed for the current polygon and the data is transferred to the frame buffer.
At this point, there are no other pixels on the current scan line which contain a portion of the polygon. This is detected by the scan line logic which indicates that the data contained in the intermediate scan line buffer is composited into the frame buffer and processing is to continue at the next scan line.
The next scan line, illustrated by FIG. 13f, consists of two edges, each intersecting a separate pixel. Each edge is processed individually to retrieve the corresponding sub-pixel masks (illustrated by FIG. 13g) which are stored in the intermediate scan line buffer and subsequently composited into the frame buffer.
The third scan line to be processed is illustrated by FIG. 13h. The sub-pixel masks for the three pixels are determined according to the process described with respect to the first two scan lines. However, it should be noted that the pixel at location (1.0, 2.0) involves two separate edges. Provisions are made in the scan line logic to provide for this case. When the first edge intersects this pixel, the scan line logic detects that the pixel location in the intermediate scan line buffer has not been written into (with respect to the current scan line). Thus, the sub-pixel mask is written into the intermediate scan line buffer. When the second edge which intersects the pixel is processed, the logic detects that the particular pixel in intermediate scan line buffer has been written into and combines the sub-pixel mask with the sub-pixel mask in the intermediate scan line buffer by logically ANDing the two sub-pixel masks together (this is illustrated by FIG. 13i).
During the processing of the rightmost pixel, the edge logic determines that the extrapolated right edge of the pixel extends beyond the second vertex. This indicates that the right edge is complete and the right edge is deleted from the active edge list. The sub-pixel masks for the edges which form the second vertex are individually determined and combined (as illustrated by FIG. 13j) by logically ANDing the two sub-pixel masks together. The resultant coverage for the scan line is illustrated by FIG. 13k.
The next scan line, illustrated by FIG. 13l, consists of the third vertex of the polygon. The edges which form the vertex are extrapolated to determine the intersection points (i.e., the edge entry and exit points of the pixel). By comparison of the locations of the intersection points with the location of the vertex, it is determined that the right edge of the pixel, determined by extrapolating the right edge of the vertex, extends beyond the left edge of the vertex and the left edge of the pixel, determined by extrapolating the left edge of the vertex, extends beyond the right edge of the vertex. Thus, the left and right edges are removed from the active edge list (the active edge list is now empty) and, as illustrated by FIG. 13m, the sub-pixel masks for each edge which forms the vertex are retrieved from the coverage table, logically ANDed together and stored in the intermediate scan line buffer (illustrated by FIG. 13n). The sub-pixel masks stored in the intermediate scan line buffer are subsequently composited with the corresponding pixels in the frame buffer to complete processing of the polygon.
Referring again to FIG. 6, once the correct partial coverage of boundary pixels or edges of the polygon are determined, at block 310, the pixels between edge pixels are filled. Preferably as each edge pixel is computed, an x value is identified with it which represents the first non-edge pixel interior of the polygon. This pixel is the first pixel to be filled inside the polygon. These x values are then sorted by y values and increasing x values and an odd/even rule is used to begin a fill span and end a fill span for any number of spans on a current scan line.
The first x value in the sorted list will be a left edge at which point the fill will begin by turning on each successive pixel. The fill process continues until a corresponding right edge is found. The list is further reviewed to see if another set of left and right edges are found in the current scan line to repeat the fill process.
Preferably, the scan line processing between edge pixels is accomplished using a simple linear interpolation loop; for example for a planar polygon:
while (x<right-- x)
process-- pixel (x, y, z, r, g, b)
b+=dbDx where x, y, z denote the display coordinate value, with z denoting depth, and r, g, b denote the color coordinate values red, green, blue, the x, y, z, r, g, b are the x, y, z, r, g, b values to be computed, "right-- x" is the right edge pixel of the span being processed, "scale-- x" is the pixel increment across the scan line, "+=" represents the operation of incrementing the variable to the left of the operator by the variable to the right of the operator and dz Dx, dr Dx, dg Dx and db Dx respectively are the slopes of z, r, g, b with respect to x and are computed according to the following equations (assuming the edge pixels are identified by the end points of left edge P1-P3 and the end points of right edge P1-P2 as shown in FIG. 14): ##EQU2## where dx1, dy1, dz1, dr1, dg1 and db1 are the difference values for the x, y, z, r, g, b values between the end points P1, P2 which form the right edge of the span and dx3, dy3, dz3, dr3, dg3 and db3 are the difference values for the x, y, z, r, g, b values between the end points P1, P3 which form the left edge of the span on a scan line.
Once a scan line of polygon has been converted to pixel values, at block 320, FIG. 6, the z or depth value of each pixel of the polygon (hereinafter referred to as za) is compared to the z buffer value of each background pixel (zb) to be composited with the polygon to determine the set of compositing equations to use. By employing the z value of each pixel as a control for determining the compositing equations to use, an accurate rendering of the polygon composited with the background can be achieved.
At block 330, if za is less zb, then the polygon is in front of the background pixel and at block 340 the pixel of the polygon is composited with the background pixel.
FIG. 15a sets forth the compositing equations employed when a single alpha value is used to describe the extent of coverage of a pixel and FIG. 15b sets forth the equations employed to composite a pixel to the background pixel when sub-pixel masks are used to represent the coverage at a pixel.
Thus, if za is less than zb, and the alpha value is used to describe coverage, the color of the background pixel upon compositing is computed according to the equations 540, 545, 550, FIG. 15a, the alpha value for the background pixel is updated according to equation 555 and the z buffer value is set, according to FIG. 560, to equal the z value of the pixel of the polygon (za).
If sub-pixel masks are used to describe coverage, referring to FIG. 15b, the coverage value αa is determined from equations 650 and 655 and is simply determined according to the number of sub-pixel elements in the mask that are set to 1 divided by the total number of sub-pixels in the mask. Equations 660 and 665 determines the color composite of the polygon pixel and the background pixel which is the updated color of the background pixel. Equation 670 derives the resultant sub-pixel mask to be stored in the α channel of the background pixel. The sub-pixel mask is the logical OR of the previously stored sub-pixel mask of the background pixel and the mask of the pixel of the polygon. In as much as the new polygon is closer to the viewer, the z buffer value is updated to be the z value of the polygon pixel (equation 675).
If the background pixel is in front of the polygon pixel, block 350, FIG. 6, and if a single alpha value is used to describe the extent of coverage of a pixel, at block 360, FIG. 6, the background pixel and polygon pixel are composited and the alpha value is updated. Equations 565, 570, 575, FIG. 15a, are used to determine the color of the background pixel upon compositing and equation 580 is used to determine the updated value of alpha. The z buffer value remains the same because the polygon pixel composited with the background is behind the background pixel (equation 585).
At block 350, FIG. 6, if masks are used to describe coverage, if zb is less than za, the background pixel is in front of the polygon pixel and at block 360 the polygon pixel is composited with the background pixel using the equations 680, 685, 690, 695, 700, 705 and 710 in FIG. 15b. These equations are employed when the z value of the polygon pixel is greater than that of the background pixel and the mask is not already full, that is, the pixel has not been already occluded by a polygon. The sub-mask used to derive the proportional color contribution to a pixel is first logically ANDed with the logical complement of the stored background sub-pixel mask as shown in equation 680. This operation filters out pixel areas which could not possibly be visible as indicated by the mask of the background pixel. The result of the composite mask operation are those sub-pixels which could be visible, thus contributing to the coverage value (equation 685). The pixel color is determined by equation 690 and the resultant sub-pixel mask and color of the background pixel are determined by equations 695, 700 and 705. The value of the z buffer is not changed because the background pixel is in front of the polygon pixel (equation 710.
At block 370, FIG. 6, if za equals zb, two methods may be employed to perform the composition operation of block 380. The first method is to simply over write the background pixel color with the color of the polygon pixel. This method is used when, with respect to the case where the alpha value represents coverage, the sum of the alpha values is greater than or equal to the value of one. Similarly, this method is employed when the logical OR of the stored mask and the polygon sub-pixel mask result in a full mask (that is, all bits in the mask are set to 1). When the alpha value is greater than or equal to a value of one or when the mask is full it is unclear, given the stored information, what compositing had previously occurred. Because statistics have shown that the incoming information is typically of greater importance than the background information, the incoming information is used to determine the color of the pixel.
Equations 590, 595, 600, 605 and 610 are used to determine the color and alpha value z buffer value upon compositing and equations 715, 720, 725, 730, 735, and 740, FIG. 15b, are utilized when sub-pixel masks are used.
If the summed alpha value is less than 1 or the mask is not full when the background pixel mask and the polygon mask are logically ORed together, that is, (Ma|Mb)≠ to full mask, there is a strong likelihood that the pixel contains a vertex pixel and a simple over write will result in an undesirable ripple along the composited edge. Therefore, at block 390, FIG. 6, it is preferred that the compositing equations, as indicated by equations 615, 620, 625, 630 and 365, FIG. 15a, for the alpha value approach, and equations 745, 750, 755, 760, 765 and 770, FIG. 15b, for the mask approach, are used in order that the smoothness of the anti-aliased edge is preserved.
FIGS. 16a, 16b, 16c, 16d, 16e, 16f, 16g, 16h, and 16i and FIGS. 17a, 17b and 17c illustrate in detail the depth ordered dependence error and the improvement the system of the present invention achieves by minimizing the depth order dependence error while maintaining the system to be a single pass system.
In this illustration, three facets of polygons at a particular pixel are used (see FIG. 16a). The final color of the pixel output to the display device is dependent upon both the depth order (e.g. zA<zB<zC) and time order of rendering (e.g. tA<tB<tC). FIGS. 16b, 16c, 16d, 16e, 16f, 16g, 16h, and 16i illustrate the resulting color for different combinations of depth orders and time orders of rendering. The value zA, zB and zC respectively represent the z or depth value for facets A, B and C. Similarly, tA, tB and tC respectively represent the time order of rendering facet A, facet B and facet C.
The tables depicted in FIGS. 17a, 17b and 17c further illustrate the effectiveness of the system of the present invention in minimizing the depth ordered dependence error while maintaining a fast single pass system. The table of FIG. 17a illustrates the ideal color values to be rendered given an infinite amount of time to render the polygons properly, preferably through a multiple pass system. FIG. 17b illustrates the color values rendered using the system of the present invention. FIG. 17c provides the error analysis between the ideal color values and the color values rendered according to the system of the present invention. As can be observed from FIG. 17c, the error margin, on the whole, is minimal and is far outweighed by the time saved rendering anti-aliased polygons in a single pass system.
While the invention has been described in conjunction with the preferred embodiment, it is evident that numerous alternatives modifications, variations and uses will be apparent to those skilled in the art in light of the foregoing description.
|Cited Patent||Filing date||Publication date||Applicant||Title|
|US4684990 *||12 Apr 1985||4 Aug 1987||Ampex Corporation||Method and apparatus for combining multiple video images in three dimensions|
|US4704605 *||17 Dec 1984||3 Nov 1987||Edelson Steven D||Method and apparatus for providing anti-aliased edges in pixel-mapped computer graphics|
|US4780711 *||12 Apr 1985||25 Oct 1988||International Business Machines Corporation||Anti-aliasing of raster images using assumed boundary lines|
|US4825391 *||20 Jul 1987||25 Apr 1989||General Electric Company||Depth buffer priority processing for real time computer image generating systems|
|US4873515 *||16 Oct 1987||10 Oct 1989||Evans & Sutherland Computer Corporation||Computer graphics pixel processing system|
|US4908780 *||14 Oct 1988||13 Mar 1990||Sun Microsystems, Inc.||Anti-aliasing raster operations utilizing sub-pixel crossing information to control pixel shading|
|US4918626 *||9 Dec 1987||17 Apr 1990||Evans & Sutherland Computer Corp.||Computer graphics priority system with antialiasing|
|US4992780 *||29 Sep 1988||12 Feb 1991||U.S. Philips Corporation||Method and apparatus for storing a two-dimensional image representing a three-dimensional scene|
|Citing Patent||Filing date||Publication date||Applicant||Title|
|US5339386 *||8 Aug 1991||16 Aug 1994||Bolt Beranek And Newman Inc.||Volumetric effects pixel processing|
|US5343558 *||19 Feb 1991||30 Aug 1994||Silicon Graphics, Inc.||Method for scan converting shaded triangular polygons|
|US5347618 *||3 Jun 1993||13 Sep 1994||Silicon Graphics, Inc.||Method for display rendering by determining the coverage of pixels in polygons|
|US5377313 *||29 Jan 1992||27 Dec 1994||International Business Machines Corporation||Computer graphics display method and system with shadow generation|
|US5381522 *||20 Mar 1992||10 Jan 1995||Canon Kabushiki Kaisha||Image processing apparatus and method|
|US5402533 *||22 Apr 1993||28 Mar 1995||Apple Computer, Inc.||Method and apparatus for approximating a signed value between two endpoint values in a three-dimensional image rendering device|
|US5404431 *||18 Nov 1992||4 Apr 1995||Ricoh Company, Ltd.||Image drawing with improved area ratio approximation process|
|US5416893 *||26 Oct 1993||16 May 1995||Hewlett-Packard Co.||Method for displaying edging or decaling of an object in a graphics display|
|US5461703 *||13 Oct 1992||24 Oct 1995||Hewlett-Packard Company||Pixel image edge enhancement method and system|
|US5515484 *||6 Oct 1993||7 May 1996||Silicon Graphics, Inc.||Method and apparatus for rendering volumetric images|
|US5517603 *||19 Dec 1994||14 May 1996||Apple Computer, Inc.||Scanline rendering device for generating pixel values for displaying three-dimensional graphical images|
|US5526474 *||22 Oct 1992||11 Jun 1996||Ricoh Company, Ltd.||Image drawing with improved process for area ratio of pixel|
|US5583974 *||4 May 1994||10 Dec 1996||Apple Computer, Inc.||Computer graphics system having high performance multiple layer Z-buffer|
|US5592601 *||30 Sep 1994||7 Jan 1997||Apple Computer, Inc.||Method and apparatus for incremental acceleration of the rendering process utilizing multiple scanline rendering devices|
|US5606650 *||7 Jun 1995||25 Feb 1997||Apple Computer, Inc.||Method and apparatus for storage and retrieval of a texture map in a graphics display system|
|US5625772 *||7 Aug 1992||29 Apr 1997||Hitachi, Ltd.||Gray-scale font generating apparatus utilizing a blend ratio|
|US5640496 *||4 Feb 1991||17 Jun 1997||Medical Instrumentation And Diagnostics Corp. (Midco)||Method and apparatus for management of image data by linked lists of pixel values|
|US5649078 *||22 Mar 1995||15 Jul 1997||International Business Machines Corporation||Efficient two-pass rasterization scheme utilizing visibility information|
|US5684939 *||8 Aug 1996||4 Nov 1997||Silicon Graphics, Inc.||Antialiased imaging with improved pixel supersampling|
|US5706415 *||20 Dec 1991||6 Jan 1998||Apple Computer, Inc.||Method and apparatus for distributed interpolation of pixel shading parameter values|
|US5724561 *||3 Nov 1995||3 Mar 1998||3Dfx Interactive, Incorporated||System and method for efficiently determining a fog blend value in processing graphical images|
|US5739826 *||6 May 1996||14 Apr 1998||Ricoh Company, Ltd.||Polygon display based on x coordinates of edges on scan line|
|US5742277 *||6 Oct 1995||21 Apr 1998||Silicon Graphics, Inc.||Antialiasing of silhouette edges|
|US5757375 *||25 Jul 1995||26 May 1998||International Business Machines Corporation||Computer graphics system and method employing frame buffer having subpixel field, display fields and a control field for relating display fields to the subpixel field|
|US5808627 *||22 Apr 1994||15 Sep 1998||Apple Computer, Inc.||Method and apparatus for increasing the speed of rendering of objects in a display system|
|US5841447 *||2 Aug 1995||24 Nov 1998||Evans & Sutherland Computer Corporation||System and method for improving pixel update performance|
|US5860925 *||27 Jun 1997||19 Jan 1999||Siemens Medical Systems, Inc.||Ultrasound scan conversion method|
|US5872902 *||19 Jun 1995||16 Feb 1999||Nihon Unisys, Ltd.||Method and apparatus for rendering of fractional pixel lists for anti-aliasing and transparency|
|US5920687 *||7 Jun 1995||6 Jul 1999||Apple Computer, Inc.||Z-buffer storage based on opacity and depth using pointers|
|US5940080 *||12 Sep 1996||17 Aug 1999||Macromedia, Inc.||Method and apparatus for displaying anti-aliased text|
|US5977987 *||26 Jul 1996||2 Nov 1999||Raycer, Incorporated||Method and apparatus for span and subspan sorting rendering system|
|US6057855 *||2 Jul 1997||2 May 2000||Hewlett-Packard Company||Method and apparatus for providing polygon pixel sub-sample information using incremental means|
|US6072500 *||17 Sep 1997||6 Jun 2000||Silicon Graphics, Inc.||Antialiased imaging with improved pixel supersampling|
|US6104407 *||23 Sep 1997||15 Aug 2000||Ati Technologies, Inc.||Method and apparatus for processing fragment pixel information in a three-dimensional graphics processing system|
|US6115049 *||30 Sep 1996||5 Sep 2000||Apple Computer, Inc.||Method and apparatus for high performance antialiasing which minimizes per pixel storage and object data bandwidth|
|US6157393 *||15 Jul 1999||5 Dec 2000||Intergraph Corporation||Apparatus and method of directing graphical data to a display device|
|US6181355||15 Jul 1999||30 Jan 2001||3Dlabs Inc. Ltd.||Graphics processing with transcendental function generator|
|US6188410||15 Jul 1999||13 Feb 2001||3Dlabs Inc. Ltd.||System for processing vertices from a graphics request stream|
|US6429877||30 Jul 1999||6 Aug 2002||Hewlett-Packard Company||System and method for reducing the effects of aliasing in a computer graphics system|
|US6459453||15 Jul 1999||1 Oct 2002||3Dlabs Inc. Ltd.||System for displaying a television signal on a computer monitor|
|US6476807 *||20 Aug 1999||5 Nov 2002||Apple Computer, Inc.||Method and apparatus for performing conservative hidden surface removal in a graphics processor with deferred shading|
|US6476816||15 Jul 1999||5 Nov 2002||3Dlabs Inc. Ltd.||Multi-processor graphics accelerator|
|US6480913||15 Jul 1999||12 Nov 2002||3Dlabs Inc. Led.||Data sequencer with MUX select input for converting input data stream and to specific output data stream using two exclusive-or logic gates and counter|
|US6518971||15 Jul 1999||11 Feb 2003||3Dlabs Inc. Ltd.||Graphics processing system with multiple strip breakers|
|US6525737||20 Aug 1999||25 Feb 2003||Apple Computer, Inc.||Graphics processor with pipeline state storage and retrieval|
|US6545686||2 Feb 1999||8 Apr 2003||Oak Technology, Inc.||Cache memory and method for use in generating computer graphics texture|
|US6552723||20 Aug 1999||22 Apr 2003||Apple Computer, Inc.||System, apparatus and method for spatially sorting image data in a three-dimensional graphics pipeline|
|US6577305||20 Aug 1999||10 Jun 2003||Apple Computer, Inc.||Apparatus and method for performing setup operations in a 3-D graphics pipeline using unified primitive descriptors|
|US6577307||20 Sep 1999||10 Jun 2003||Silicon Integrated Systems Corp.||Anti-aliasing for three-dimensional image without sorting polygons in depth order|
|US6577316||15 Jul 1999||10 Jun 2003||3Dlabs, Inc., Ltd||Wide instruction word graphics processor|
|US6577317||20 Aug 1999||10 Jun 2003||Apple Computer, Inc.||Apparatus and method for geometry operations in a 3D-graphics pipeline|
|US6591020||23 Dec 1998||8 Jul 2003||Xerox Corporation||Antialiazed high-resolution frame buffer architecture|
|US6597363||20 Aug 1999||22 Jul 2003||Apple Computer, Inc.||Graphics processor with deferred shading|
|US6614444||20 Aug 1999||2 Sep 2003||Apple Computer, Inc.||Apparatus and method for fragment operations in a 3D-graphics pipeline|
|US6618185||28 Nov 2001||9 Sep 2003||Micronic Laser Systems Ab||Defective pixel compensation method|
|US6636232 *||12 Jan 2001||21 Oct 2003||Hewlett-Packard Development Company, L.P.||Polygon anti-aliasing with any number of samples on an irregular sample grid using a hierarchical tiler|
|US6664959||29 Aug 2002||16 Dec 2003||Apple Computer, Inc.||Method and apparatus for culling in a graphics processor with deferred shading|
|US6674440||23 Mar 2000||6 Jan 2004||3Dlabs, Inc., Inc. Ltd.||Graphics processor for stereoscopically displaying a graphical image|
|US6693639||7 Nov 2002||17 Feb 2004||Apple Computer, Inc.||Graphics processor with pipeline state storage and retrieval|
|US6697064||8 Jun 2001||24 Feb 2004||Nvidia Corporation||System, method and computer program product for matrix tracking during vertex processing in a graphics pipeline|
|US6717576||20 Aug 1999||6 Apr 2004||Apple Computer, Inc.||Deferred shading graphics pipeline processor having advanced features|
|US6771264||17 Dec 1998||3 Aug 2004||Apple Computer, Inc.||Method and apparatus for performing tangent space lighting and bump mapping in a deferred shading graphics processor|
|US6813062 *||8 Sep 2003||2 Nov 2004||Micronic Laser Systems Ab||Defective pixel compensation method|
|US6833836 *||23 Jul 2002||21 Dec 2004||Sony Computer Entertainment, Inc.||Image rendering process for displaying three-dimensional image information on a two-dimensional screen|
|US6919906 *||8 May 2001||19 Jul 2005||Microsoft Corporation||Discontinuity edge overdraw|
|US6965119||10 Sep 2002||15 Nov 2005||Micronic Laser Systems Ab||Method and apparatus of calibrating multi-position SLM elements|
|US6982718||30 Nov 2001||3 Jan 2006||Nvidia Corporation||System, method and computer program product for programmable fragment processing in a graphics pipeline|
|US7006101||25 Jun 2002||28 Feb 2006||Nvidia Corporation||Graphics API with branching capabilities|
|US7061226||14 Jan 2004||13 Jun 2006||Micronic Laser Systems Ab||Method to detect a defective element|
|US7142211 *||7 Jul 2003||28 Nov 2006||Arcsoft, Inc.||Graphic engine for fill style transferring in a resource-constrained device|
|US7158280||14 Nov 2005||2 Jan 2007||Micronic Laser Systems Ab||Methods and systems for improved boundary contrast|
|US7164426||28 Nov 2000||16 Jan 2007||Apple Computer, Inc.||Method and apparatus for generating texture|
|US7167181||9 Jun 2003||23 Jan 2007||Apple Computer, Inc.||Deferred shading graphics pipeline processor having advanced features|
|US7170513||1 Jul 2002||30 Jan 2007||Nvidia Corporation||System and method for display list occlusion branching|
|US7280121||23 Sep 2004||9 Oct 2007||Sony Corporation||Image processing apparatus and method of same|
|US7286133||10 May 2005||23 Oct 2007||Nvidia Corporation||System, method and computer program product for programmable fragment processing|
|US7286138||22 Feb 2005||23 Oct 2007||Microsoft Corporation||Discontinuity edge overdraw|
|US7432937 *||30 Jun 2005||7 Oct 2008||Intel Corporation||System and method for concave polygon rasterization|
|US7433191||30 Sep 2005||7 Oct 2008||Apple Inc.||Thermal contact arrangement|
|US7456838||18 Feb 2005||25 Nov 2008||Nvidia Corporation||System and method for converting a vertex program to a binary format capable of being executed by a hardware graphics pipeline|
|US7486337 *||22 Dec 2003||3 Feb 2009||Intel Corporation||Controlling the overlay of multiple video signals|
|US7518616||15 Jul 1999||14 Apr 2009||3Dlabs, Inc. Ltd.||Graphics processor with texture memory allocation system|
|US7538779 *||13 Jul 2006||26 May 2009||Luxology, Llc||Method of rendering pixel images from abstract datasets|
|US7577930||23 Jun 2005||18 Aug 2009||Apple Inc.||Method and apparatus for analyzing integrated circuit operations|
|US7598711||23 Nov 2005||6 Oct 2009||Apple Inc.||Power source switchover apparatus and method|
|US7599044||23 Jun 2005||6 Oct 2009||Apple Inc.||Method and apparatus for remotely detecting presence|
|US7616200||10 Jun 1999||10 Nov 2009||3Dlabs Inc. Ltd.||System for reducing aliasing on a display device|
|US7755634||22 Nov 2005||13 Jul 2010||Nvidia Corporation||System, method and computer program product for branching during programmable vertex processing|
|US7764287||22 May 2006||27 Jul 2010||Arcsoft, Inc.||Graphic engine for approximating a quadratic bezier curve in a resource-constrained device|
|US7808503||19 Dec 2006||5 Oct 2010||Apple Inc.||Deferred shading graphics pipeline processor having advanced features|
|US7891818||12 Dec 2007||22 Feb 2011||Evans & Sutherland Computer Corporation||System and method for aligning RGB light in a single modulator projector|
|US8077378||12 Nov 2009||13 Dec 2011||Evans & Sutherland Computer Corporation||Calibration system and method for light modulation device|
|US8155826||13 Mar 2008||10 Apr 2012||Aisin Aw Co., Ltd.||Vehicle behavior learning apparatuses, methods, and programs|
|US8184861||11 Mar 2008||22 May 2012||Aisin Aw Co., Ltd.||Feature information management apparatuses, methods, and programs|
|US8358317||26 May 2009||22 Jan 2013||Evans & Sutherland Computer Corporation||System and method for displaying a planar image on a curved surface|
|US8410977 *||17 Feb 2011||2 Apr 2013||Honeywell International Inc.||Methods and systems for identifying hazardous flight zone areas on a display|
|US8702248||11 Jun 2009||22 Apr 2014||Evans & Sutherland Computer Corporation||Projection method for reducing interpixel gaps on a viewing surface|
|US8825353||10 Sep 2008||2 Sep 2014||Aisin Aw Co., Ltd.||Parking support systems, parking support methods, and parking support programs|
|US8860742 *||2 May 2012||14 Oct 2014||Nvidia Corporation||Coverage caching|
|US8892517||8 Jun 2007||18 Nov 2014||Aisin Aw Co., Ltd.||Data updating system, navigation device, server, and method of data updating|
|US9218641 *||19 Jun 2012||22 Dec 2015||Exelis, Inc.||Algorithm for calculating high accuracy image slopes|
|US9298311||23 Jun 2005||29 Mar 2016||Apple Inc.||Trackpad sensitivity compensation|
|US9479661 *||4 Jun 2015||25 Oct 2016||Brother Kogyo Kabushiki Kaisha||Image-reading device detecting edge of sheet by setting mask region including part of lines|
|US9531896||4 Jun 2015||27 Dec 2016||Brother Kogyo Kabushiki Kaisha||Image-reading device detecting edge of sheet by dynamically setting mask region|
|US9536019 *||4 Aug 2014||3 Jan 2017||Axis Ab||Method and system for selecting position and orientation for a monitoring camera|
|US9641826||10 Jul 2012||2 May 2017||Evans & Sutherland Computer Corporation||System and method for displaying distant 3-D stereo on a dome surface|
|US9697641 *||14 Oct 2010||4 Jul 2017||Nvidia Corporation||Alpha-to-coverage using virtual samples|
|US20020167532 *||28 Jun 2002||14 Nov 2002||Stroyan Howard D.||System and method for reducing the effects of aliasing in a computer graphics system|
|US20020196256 *||8 May 2001||26 Dec 2002||Hugues Hoppe||Discontinuity edge overdraw|
|US20030016232 *||16 Jul 2002||23 Jan 2003||Van Overveld Cornelius Wilhelmus Antonius Marie||Computer graphics system and method of rendering polygons|
|US20030081303 *||10 Sep 2002||1 May 2003||Micronic Laser Systems Ab||Method and apparatus using an SLM|
|US20040047023 *||8 Sep 2003||11 Mar 2004||Micronic Laser Systems Ab||Defective pixel compensation method|
|US20040130552 *||9 Jun 2003||8 Jul 2004||Duluk Jerome F.||Deferred shading graphics pipeline processor having advanced features|
|US20040207386 *||14 Jan 2004||21 Oct 2004||Micronic Laser Systems Ab||Method to detect a defective element|
|US20050007365 *||7 Jul 2003||13 Jan 2005||Jiangen Cao||Graphic engine for rasterizing a straight edge in a resource-constrained device|
|US20050007366 *||7 Jul 2003||13 Jan 2005||Shuijun Lu||Graphic engine for fill style transferring in a resource-constrained device|
|US20050068333 *||23 Sep 2004||31 Mar 2005||Teruyuki Nakahashi||Image processing apparatus and method of same|
|US20050104893 *||22 Sep 2004||19 May 2005||Sharp Kabushiki Kaisha||Three dimensional image rendering apparatus and three dimensional image rendering method|
|US20050134605 *||22 Feb 2005||23 Jun 2005||Microsoft Corporation||Discontinuity edge overdraw|
|US20050134739 *||22 Dec 2003||23 Jun 2005||Bian Qixiong J.||Controlling the overlay of multiple video signals|
|US20050259103 *||10 May 2005||24 Nov 2005||Nvidia Corporation||System, method and computer program product for programmable fragment processing|
|US20060077506 *||14 Nov 2005||13 Apr 2006||Micronic Laser Systems Ab||Methods and systems for improved boundary contrast|
|US20060197509 *||1 Mar 2005||7 Sep 2006||Takashi Kanamori||Method and apparatus for voltage regulation|
|US20060227349 *||5 Apr 2006||12 Oct 2006||Aisin Aw Co., Ltd.||Systems, methods, and programs for rotating font data and map display systems, methods, and programs that display rotated font data|
|US20060256115 *||22 May 2006||16 Nov 2006||Arcsoft, Inc.||Graphic Engine for Approximating a Quadratic Bezier Curve in a Resource-Constrained Device|
|US20060290677 *||23 Jun 2005||28 Dec 2006||Lyon Benjamin B||Trackpad sensitivity compensation|
|US20070002074 *||30 Jun 2005||4 Jan 2007||Bimal Poddar||System and method for concave polygon rasterization|
|US20070024639 *||13 Jul 2006||1 Feb 2007||Luxology, Llc||Method of rendering pixel images from abstract datasets|
|US20070076378 *||30 Sep 2005||5 Apr 2007||Blanco Richard L Jr||Thermal contact arrangement|
|US20070114968 *||23 Nov 2005||24 May 2007||Krah Christoph H||Power source switchover apparatus and method|
|US20070165035 *||19 Dec 2006||19 Jul 2007||Apple Computer, Inc.||Deferred shading graphics pipeline processor having advanced features|
|US20080240506 *||11 Mar 2008||2 Oct 2008||Aisin Aw Co., Ltd.||Feature information management apparatuses, methods, and programs|
|US20080243312 *||13 Mar 2008||2 Oct 2008||Aisin Aw Co., Ltd.||Vehicle behavior learning apparatuses, methods, and programs|
|US20090088959 *||10 Sep 2008||2 Apr 2009||Aisin Aw Co., Ltd.||Parking support systems, parking support methods, and parking support programs|
|US20090177706 *||8 Jun 2007||9 Jul 2009||Aisin Aw Co., Ltd.||Data Updating System, Navigation Device, Server, and Method of Data Updating|
|US20110090250 *||14 Oct 2010||21 Apr 2011||Molnar Steven E||Alpha-to-coverage using virtual samples|
|US20110183301 *||27 Jan 2010||28 Jul 2011||L-3 Communications Corporation||Method and system for single-pass rendering for off-axis view|
|US20120212367 *||17 Feb 2011||23 Aug 2012||Honeywell International Inc.||Methods and systems for identifying hazardous flight zone areas on a display|
|US20120280992 *||1 May 2012||8 Nov 2012||Shebanow Michael C||Grid walk sampling|
|US20120281004 *||2 May 2012||8 Nov 2012||Shebanow Michael C||Coverage caching|
|US20150046129 *||4 Aug 2014||12 Feb 2015||Axis Ab||Method and system for selecting position and orientation for a monitoring camera|
|US20150373218 *||4 Jun 2015||24 Dec 2015||Brother Kogyo Kabushiki Kaisha||Image-Reading Device Detecting Edge of Sheet by Setting Mask Region Including Part of Lines|
|USRE38078||21 Jan 1999||15 Apr 2003||Apple Computer, Inc.||Graphical rendering system using simultaneous parallel query Z-buffer and method therefor|
|CN1129091C *||7 Oct 1996||26 Nov 2003||任天堂株式会社||Antialiasing method and system of silhouette edges|
|DE19827726B4 *||22 Jun 1998||24 Feb 2005||Hewlett-Packard Co. (N.D.Ges.D.Staates Delaware), Palo Alto||Verfahren und Vorrichtung zum Liefern von Polygonpixelunterabtastinformationen unter Verwendung einer Inkrementaleinrichtung|
|EP0767444A2 *||7 Oct 1996||9 Apr 1997||Nintendo Co., Ltd.||Antialiasing of silhouette edges|
|EP0767444A3 *||7 Oct 1996||11 Feb 1998||Nintendo Co., Ltd.||Antialiasing of silhouette edges|
|EP0910044A2 *||24 Sep 1998||21 Apr 1999||Digital Equipment Corporation||Method and apparatus for compositing colors of images with memory constraints|
|EP0910044A3 *||24 Sep 1998||3 May 2000||Digital Equipment Corporation||Method and apparatus for compositing colors of images with memory constraints|
|EP1519319A2 *||21 Sep 2004||30 Mar 2005||Sony Corporation||Method and apparatus for antialiased blending in image rendering|
|EP1519319A3 *||21 Sep 2004||10 May 2006||Sony Corporation||Method and apparatus for antialiased blending in image rendering|
|WO1997005575A1 *||26 Jul 1996||13 Feb 1997||Raycer, Incorporated||Method and apparatus for span sorting rendering system|
|WO1997005576A1 *||26 Jul 1996||13 Feb 1997||Silicon Engines, Inc.||Method and apparatus for span and subspan sorting rendering system|
|WO1999064990A2 *||10 Jun 1999||16 Dec 1999||Intergraph Corporation||System for reducing aliasing on a display device|
|WO1999064990A3 *||10 Jun 1999||13 Apr 2000||Intergraph Corp||System for reducing aliasing on a display device|
|WO2000019372A1 *||28 Sep 1998||6 Apr 2000||Orbotech Ltd.||Pixel coding and image processing method|
|WO2000065540A1 *||20 Apr 2000||2 Nov 2000||Broadcom Corporation||Method and system for providing implicit edge antialiasing|
|WO2002101656A1 *||4 Jun 2002||19 Dec 2002||Nvidia Corporation||System, method and computer program product for matrix tracking during vertex processing in a graphics pipeline|
|WO2003010718A1 *||24 Jun 2002||6 Feb 2003||Koninklijke Philips Electronics N.V.||Computer graphics system and method of rendering polygons|
|U.S. Classification||345/421, 345/422, 345/443, 345/629|
|International Classification||G06T15/50, G09G5/36, G06T11/60, G06T11/20|
|19 Mar 1990||AS||Assignment|
Owner name: SUN MICROSYSTEMS, INC., 2550 GARCIA AVENUE, MOUNTA
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST.;ASSIGNORS:WELLS, STUART C.;LOO, JAMES V.;WALLNER, DAWN M.;REEL/FRAME:005263/0628
Effective date: 19900316
|22 Sep 1992||CC||Certificate of correction|
|13 Dec 1995||FPAY||Fee payment|
Year of fee payment: 4
|9 Sep 1997||CC||Certificate of correction|
|15 Dec 1999||FPAY||Fee payment|
Year of fee payment: 8
|31 Dec 2003||REMI||Maintenance fee reminder mailed|
|16 Jun 2004||LAPS||Lapse for failure to pay maintenance fees|
|10 Aug 2004||FP||Expired due to failure to pay maintenance fee|
Effective date: 20040616